Compression of 16 bit data using predictor values

Abstract

Systems and methods compress and decompress 16 bit data. The 16 bit data may be signed or unsigned and represented in a fixed point or floating point format. A fixed block size of data is compressed into a fixed length format. Data compressed using a medium quality compression scheme may be efficiently decompressed in hardware. Data may be efficiently compressed and decompressed in hardware using a high quality compression scheme. The high quality compression scheme has a lower compression ratio compared with the medium quality compression scheme, but is near lossless in terms of quality.

1. A computer-implemented method of compressing a plurality of 16-bit values stored in a memory into a compressed format, comprising:
 transforming the plurality of 16-bit values to produce four predictor values and a plurality of error values for transmission to a quantization unit; 
 quantizing the four predictor values based on a maximum number of bits needed to represent three of the four predictor values to produce quantized predictor values for transmission to an encoding unit; 
 quantizing the plurality of error values based on a maximum number of bits needed to represent each one of the plurality of error values to produce quantized error values for transmission to the encoding unit; 
 converting the quantized predictor values and the quantized error values to nonnegative quantized predictor values and nonnegative quantized error values; and 
 encoding the nonnegative quantized predictor values and nonnegative quantized error values into the compressed format. 
 
     
     
       2. The method of  claim 1 , wherein the plurality of 16-bit values are represented in a fixed point data format. 
     
     
       3. The method of  claim 2 , wherein the fixed point data format includes a sign. 
     
     
       4. The method of  claim 1 , wherein the plurality of 16-bit values are represented in a floating point data format. 
     
     
       5. The method of  claim 1 , further comprising the steps of:
 computing control information indicating the maximum number of bits needed to represent the predictor values and the maximum number of bits needed to represent each one of the plurality of error values; and 
 encoding the control information into the compressed format. 
 
     
     
       6. The method of  claim 1 , wherein twelve error values are included in the plurality of error values. 
     
     
       7. The method of  claim 1 , wherein one of the four predictor values is a mean of the plurality of 16-bit values. 
     
     
       8. The method of  claim 1 , wherein sixteen 16-bit values are represented by the compressed format. 
     
     
       9. The method of  claim 1 , wherein the quantized predictor values and the quantized error values are negative or nonnegative values. 
     
     
       10. A system for compressing a plurality of 16-bit values into a compressed format, comprising:
 a processor configured to compress the plurality of 16-bit values into the compressed format, comprising:
 a transformation unit configured to receive the 16-bit values and compute four predictor values and a plurality of error values, 
 a quantization unit configured to quantize the four predictor values and the plurality of error values to produce quantized predictor values and quantized error values, and 
 an encoding unit configured to convert the quantized predictor values and the quantized error values to nonnegative quantized predictor values and nonnegative quantized error values and compute control information and encode the nonnegative quantized predictor values and nonnegative quantized error values into the compressed format; and 
 
 a memory coupled to the processor and configured to store the plurality of 16-bit values represented in the compressed format. 
 
     
     
       11. The system of  claim 10 , wherein at least one of the 16-bit values is represented in a floating point data format. 
     
     
       12. The system of  claim 10 , wherein at least one of the 16-bit values is represented in a fixed point data format. 
     
     
       13. The system of  claim 12 , wherein the fixed point data format includes a sign. 
     
     
       14. The system of  claim 10 , wherein the control information indicates a maximum number of bits needed to represent the predictor values and a maximum number of bits needed to represent each one of the plurality of error values. 
     
     
       15. The system of  claim 10 , wherein the plurality of error values includes twelve error values corresponding to a two-dimensional block of sixteen pixels. 
     
     
       16. The system of  claim 10 , wherein one of the four predictor values is a mean of the plurality of 16-bit values. 
     
     
       17. The system of  claim 10 , wherein each of the quantized predictor values and each of the quantized error values include a sign bit. 
     
     
       18. A non-transitory computer-readable medium storing instructions that, when executed by a processor, cause the processor to compress a plurality of 16-bit values stored in a memory into a compressed format by performing the steps of:
 transforming the plurality of 16-bit values to produce four predictor values and a plurality of error values for transmission to a quantization unit; 
 quantizing the four predictor values based on a maximum number of bits needed to represent three of the four predictor values to produce quantized predictor values for transmission to an encoding unit; 
 quantizing the plurality of error values based on a maximum number of bits needed to represent each one of the plurality of error values to produce quantized error values for transmission to the encoding unit; 
 converting the quantized predictor values and the quantized error values to nonnegative quantized predictor values and nonnegative quantized error values; and 
 encoding the nonnegative quantized predictor values and the nonnegative quantized error values into the compressed format. 
 
     
     
       19. The computer-readable medium of  claim 18 , further comprising the steps of:
 computing control information indicating the maximum number of bits needed to represent the predictor values and the maximum number of bits needed to represent each one of the plurality of error values; and 
 encoding the control information into the compressed format. 
 
     
     
       20. The computer-readable medium of  claim 18 , wherein one of the four predictor values is a mean of the plurality of 16-bit values. 
     
     
       21. The computer-readable medium of  claim 18 , wherein sixteen 16-bit values are represented by the compressed format.